Fault diagnosis for underdetermined multistage assembly processes via an enhanced Bayesian hierarchical model

This paper presents a novel approach for optimal key characteristics-based sensor distribution in a multi-station assembly process, for the purpose of diagnosing variation sources responsible for product quality defects in a timely manner. Current approaches for sensor distribution are based on the assumption that measurement points can be allocated at arbitrary locations on the part or subassembly. This not only presents challenges in the implementation of these approaches but additionally does not allow required product assurance and quality control standards to be integrated with them, due to lack of explicit relations between measured features and geometric dimensioning and tolerancing (GD&T). Furthermore, it causes difficulty in calibration of measurement system and increases the likelihood of measurement error due to the introduction of measurement points not defined in GD&T. In the proposed approach, we develop methodology for optimal sensor allocation for 6-sigma root cause analysis that maximizes the number of measurement points placed at critical design features called Key Characteristics (KCs) which are classified into: Key Product Characteristics and Key Control Characteristics and represent critical product and process design features, respectively. In particular, KCs have defined dimensional and geometric tolerances which provides necessary design reference model for process control and diagnosis of product 6-sigma variation faults. The proposed approach allows obtaining minimum required production system 6-sigma diagnosability. A feature-based procedure is proposed which includes Genetic Algorithm-based approach (allowing pre-defined KCs as the measurement points) and state-of-the-art approaches (unrestricted location of measurement points) to iteratively include arbitrary measurement points together with KCs in the final sensor layout. A case study of automotive assembly processes is used to illustrate the proposed feature-based approach.

Dimensional control has a significant impact on the overall product quality and performance in large and complex multi-station assembly systems. From measurement data, the way to identify root causes for large variation of Key Product Characteristics (KPCs) is one of the most critical research topics in dimensional control. This paper proposes a new approach for multiple fault diagnosis in a multi-station assembly process by integrating multivariate statistical analysis with engineering model. Based on product/process information, by using the state space model, a set of fault patterns for multi-station assembly process are developed, which explicitly represent the relationship between the error sources and KPCs. The vectors of these patterns form an affine system. Afterwards, the Principal Component Analysis (PCA) is applied to conduct orthogonal diagonalization of the measurement data. Thus, the measurement data can be easily projected to the axes of the affine system. Whereby, the significance of each fault pattern shall be estimated accurately. Finally, a few case studies are also provided to validate the proposed methodology.

Root Cause Analysis of Product Service Failure Using Computer Experimentation Technique

In assembly process control and design optimization, it is critical to establish a mathematical model that describes the relationship between the dimensional quality of the final product and the various process parameters (e.g., the fixture layout and locator position deviation). This article presents a surrogate modeling methodology for multistage assembly processes to characterize the relationship between fixture layout and product dimensional quality. The mathematical structure of the model is derived from a physical analysis based on first principles and then the parameters of the model are identified using data from computer experiments. The resulting surrogate model can enable fixture layout optimization in process planning. A comprehensive case study of a multistage assembly process is also presented to demonstrate the effectiveness and high fidelity of the developed method.

Root-cause identification for quality-related problems is a key issue in quality and productivity improvement for a manufacturing process. Unfortunately, root-cause identification is also a very challenging engineering problem, particularly for a multistage manufacturing process. In this paper, root-cause identification is formulated as a problem of estimation and hypothesis testing of a general linear mixed model. First, a linear mixed fault-quality model is built to describe the cause-effect relationship between the process faults and product quality. Then, the estimation algorithms developed for a general linear mixed model are adapted to estimate the process mean and variance. Finally, a hypothesis testing method is developed to determine if process faults exist in terms of statistical significance. A detailed experimental study illustrated the effectiveness of the proposed methodology.<br><br>Note to Practitioners-Economic globalization brings intense competition among manufacturing enterprises. The key to succeed in this competitive climate is to rapidly respond to fast-changing market demands with high-quality and competitively priced products. To achieve this, we need to quickly identify root causes of quality-related problems in a complicated manufacturing system. However, the current widely adopted quality-control techniques focus more on monitoring than on root-cause identification. These techniques can efficiently detect the changes in the process but the root cause identification is often left to the plant engineers or operators. In this paper, a systematic estimation and testing method is proposed to identify the variational root causes in multistage manufacturing processes. First, a linear model is built based on the design information to describe the cause-effect relationship between the process faults and product quality. Then, an algorithm is developed to estimate the mean and variance of the process faults from the quality measurements of products. Finally, a statistical testing method is developed to determine if process faults (i.e. root causes) exist in terms of statistical significance. A detailed experimental study illustrates the effectiveness of this method. The method presented in this paper is a new quality-control technique and can be used for quality improvement of multistage manufacturing processes.

The study of fixture related errors is playing an increasingly important role in the improvement of machined part quality. In this work, the locator variability of machining fixtures is analyzed following an uncertainty approach, allowing us to estimate the effect of fixture related geometrical errors on key product characteristics of machined workpieces. Monte Carlo simulation was implemented to generate sufficient variability to determine the expected value of the measuring points of the machined workpiece surface and its related key product characteristic. The results from the Monte Carlo simulation allow sufficient range of values to construct a mathematical relationship between fixture related errors and machined workpiece key product characteristics. Furthermore, the Taguchi loss function is used to investigate the loss cost incurred by a given fixturing uncertainty scheme. An example problem of a prismatic workpiece is presented, where the locating error due to the fixturing system is evaluated in order to know its influence on the workpiece flatness and its associated uncertainty.

Joint recurrence based root cause analysis of nonlinear multivariate chemical processes

Efficient condition monitoring and fault diagnosis is an essential task to ensure the generation performance and reliability of photovoltaic (PV) systems. This paper proposes an online algorithm to diagnose faults of PV module based on multi-class support vector machine (M-SVM). The simulation models of the photovoltaic module are implemented and the output power generation characteristics of PV modules under two typical fault conditions (line-to-line fault and abnormal degradation fault) are analyzed. Through the combination of mathematical model and simulation experiments of PV modules, the fill factor (FF) and the fault type factor K are introduced as feature parameters in the fault diagnosis process. In addition, to address the nonlinear problem in fault diagnosis and improve the single support vector machine, a fault diagnosis method based on the multi-class classification method of one-against-one (OAO) algorithm is proposed. The proposed solution is implemented and simulated through experiments and the numerical result clearly demonstrates its effectiveness and diagnosis accuracy.

Batch process performance monitoring has been achieved primarily using process measurements with the extracted information being associated with the physical parameters of the process. With increasing attention now being paid to the application of on‐line real‐time process analytics through spectrometry, together with the FDA Process Analytical Technologies (PAT) initiative, the use of spectroscopic information for enhanced monitoring of reactions is gaining impetus. The harmonious integration of process data and spectroscopic data then becomes a major challenge. By integrating the process and spectroscopic measurements for multivariate statistical data modelling and analysis, it is conjectured that improved process understanding and fault diagnosis can be achieved. An investigation into combining process and spectral data using multiblock and multiresolution analysis is proposed and the results from the analysis of experimental data from two industrial application studies are presented to demonstrate the improvements achievable in terms of process performance monitoring and fault diagnosis.

Multiple phases with phase to phase transitions are important characteristics of many batch processes. The linear characteristics between phases are taken into consideration in the traditional algorithms while nonlinearities are neglected, which can lead to inaccuracy and inefficiency in monitoring. The focus of this paper is nonlinear multi-phase batch processes. A similarity metric is defined based on kernel entropy component analysis (KECA). A KECA similarity-based method is proposed for phase division and fault monitoring. First, nonlinear characteristics can be extracted in feature space via performing KECA on each preprocessed time-slice data matrix. Then phase division is achieved with the similarity variation of the extracted feature information. Then, a series of KECA models and slide-KECA models are established for steady and transitions phases respectively, which can reflect the diversity of transitional characteristics objectively and preferably deal with the stage-transition monitoring problem in multistage batch processes. Next, in order to overcome the problem that the traditional contribution plot cannot be applied to the kernel mapping space, a nonlinear contribution plot diagnosis algorithm is proposed, which is easier, more intuitive and implementable compared with the traditional one. Finally, simulations are performed on penicillin fermentation and industrial application. Specifically, the proposed method detects the abnormal agitation power and the abnormal substrate supply at 47 h and 86 h, respectively. Compared with traditional methods, it has better real-time performance and higher efficiency. Results demonstrate the ability of the proposed method to detect faults accurately and effectively in practice.

Two neural network models are compared using a process fault detection and diagnosis problem. Process fault detection and diagnosis using steady-state information is a nonlinear pattern recognition problem. In such problems the measurement pattern vectors are noisy and collection of a complete training data set is time consuming and in some cases impossible. The training data were obtained from a simulated chemical process. The process consists of a reactor and a distillation column. The dynamically capacity allocating network models were found to have the best performance as compared to backpropagation-type network models. These networks can also be used for continuously learning systems and therefore the difficulties of training data collection are avoided. >

PurposeThis paper aims to present a modeling and analysis approach for multi-station aircraft assembly to predict assembly variation. The variation accumulated in the assembly process will influence the dimensional accuracy and fatigue life of airframes. However, in digital large aircraft assembly, variation propagation analysis and modeling are still unresolved issues.Design/methodology/approachBased on an elastic structure model and variation model of multistage assembly in one station, the propagation of key characteristics, assembly reference and measurement errors are introduced. Moreover, the reposition and posture coordination are considered as major aspects. The reposition of assembly objects in a different assembly station is described using transformation and blocking of coefficient matrix in finite element equation. The posture coordination of the objects is described using homogeneous matrix multiplication. Then, the variation propagation model and analysis of large aircraft assembly are established using a discrete system diagram.FindingsThis modeling and analysis approach for multi-station aircraft assembly reveals the basic rule of variation propagation between adjacent assembly stations and can be used to predict assembly variation or potential dimension problems at a preliminary assembly phase.Practical implicationsThe modeling and analysis approaches have been used in a transport aircraft project, and the calculated results were shown to be a good prediction of variation in the actual assembly.Originality/valueAlthough certain simplifications and assumptions have been imposed, the proposed method provides a better understanding of the multi-station assembly process and creates an analytical foundation for further work on variation control and tolerance optimization.

In this paper, the problem of fault diagnosis in wind turbine will be addressed based on data-driven. The idea is to classify of data by a classifier in two classes, so that we can decide whether the state of the system is defective or not. The proposed hybrid classifier is a combining of those three approaches Bayes Statistical Algorithm, Back-propagation Neural Networks and Decision Trees. During training, the learning rate has gradually decreased in discrete steps converges to a minimum error solution of classification. The Different parts of the process were investigated, including actuators, sensors and process faults. All Process faults mainly concerned friction in the wind turbine, which might cause it damage. Our results are compared with those three approaches each separately who have formed the proposed strategy to indicate the value of our strategy. This proposed classifier has built with the "Waikato Environment for Knowledge Analysis (WEKA)" library and evaluated by observing of classification rate.

This article addresses the problem of fault diagnosis in multistation assembly systems. Fault diagnosis is to identify process faults that cause excessive dimensional variation of the product using dimensional measurements. For such problems, the challenge is solving an underdetermined system caused by a common phenomenon in practice; namely, the number of measurements is less than that of the process errors. To address this challenge, this article attempts to solve the following two problems: (i) how to utilize the temporal correlation in the time series data of each process error and (ii) how to apply prior knowledge regarding which process errors are more likely to be process faults. A novel sparse Bayesian learning method is proposed to achieve the above objectives. The method consists of three hierarchical layers. The first layer has parameterized prior distribution that exploits the temporal correlation of each process error. Furthermore, the second and third layers achieve the prior distribution representing the prior knowledge of process faults. Since posterior distributions of process faults are intractable, this article derives approximate posterior distributions via Variational Bayes inference. Numerical and simulation case studies using an actual autobody assembly process are performed to demonstrate the effectiveness of the proposed method.

Unsupervised classification of multichannel profile data using PCA: An application to an emission control system